Currently, there are a number of techniques for diagnosing problems with server systems. A first step that is taken in diagnosing a server problem is identification of the server or servers that are experiencing a problem. This may be a complex task for very large server farms including a large number of servers. A server is typically identified by placing a stick-on label on the server, which includes a server name or other identifying information. Alternatively, servers may be provided with an LCD screen or other display for displaying server identification information. For high-density server systems, these techniques are problematic for a couple of reasons. First, with stick-on labels, a user must visually inspect the servers and manually identify a server of interest, which can be a time-consuming process when a large number of servers must be inspected. Second, a frontal surface on a server for placing an LCD display is seldom available.
After identifying a particular server, the problem with the server must be diagnosed. Some servers have LCD displays for displaying diagnostic messages. Such diagnostic display subsystems can be costly, and as mentioned above, frontal surface area on a server is seldom available for such as display. And the visibility can be marginal when a server is installed in the top of an extended rack with a large number of servers. Further, the displayed information is often limited to “cryptic” text messages, usually only in the English language, due to the internal message string storage limitations, and the complexities of translation into 7 or more languages to cover worldwide needs. Such LCD displays typically display error codes that must be manually looked up in a reference manual by a maintenance person. Displayed component error codes are usually specific to particular server models, and require manual look up in a service manual for the particular server model. A picture is often needed to find the faulty component or subassembly in the server.
Other less expensive approaches have been used, such as using LED's on internal printed circuit assemblies, with very poor to no visibility. LED's have also been provided on an internal or external central panel with an appropriate plastic overlay including function names. Such display systems are costly, when costs for cables, connectors, interface chips, and printed circuit board area, are taken into account. In addition, the cables and connectors can contribute to system failure.
Server problems may also be diagnosed remotely. A first type of remote diagnostic connection is an “in-band” diagnostic connection. In-band remote diagnostic schemes typically require the server and the network to be working for server diagnoses to occur. In contrast, a second type of remote diagnostic scheme is an “out-of-band” connection, which allows diagnoses to be performed even when the server or network is non-functional. Out-of-band remote diagnostic connections have previously been implemented with a phone line and an external bus. The cabling and connectors used in such out-of-band schemes increase system costs, and can contribute to system failures. Further, the large number of cables needed for high-density server systems can make this cable-based approach problematic.
It would be desirable to provide a server diagnostic system and method based on wireless technology without many of the disadvantages found in conventional diagnostic systems and methods.